Predictors for insulin use in gestational diabetes mellitus

Area of Science:

• Obstetrics and Gynecology
• Endocrinology
• Public Health

Background:

• Gestational diabetes mellitus (GDM) affects 15% of pregnancies in Australia, with 30% requiring insulin.
• Predicting insulin needs in GDM is crucial for timely intervention.
• Limited research exists on predicting insulin therapy requirements based on GDM risk factors.

Purpose of the Study:

• To identify predictors of insulin therapy in women diagnosed with GDM.
• To establish a predictive model for insulin use in pregnant women with GDM.

Main Methods:

• Retrospective cohort study of 2,048 women with GDM (2016-2017) in Melbourne, Australia.
• Data collected from hospital records and pathology results.
• Univariable and multivariable logistic regression models used to identify predictors.

Main Results:

• Elevated fasting OGTT, previous birth weight >90th%, prior GDM diagnosis, South Asian region birth, 2hr OGTT result, BMI, and age predicted insulin use.
• The final predictive model achieved an AUC of 0.744.
• 31.6% of women with GDM required insulin therapy.

Conclusions:

• Identified key predictors for insulin therapy in GDM.
• Findings can inform counselling for newly diagnosed GDM patients.
• Highlights the utility of OGTT results and patient history in predicting insulin needs.